AWS Batch Compute Environment setup: If Minimum vCPUs is > 0 then instances keep restarting, how do I terminate them permanently?

If I change the max vCPU to 0, will the instance be launched repeatedly?
I think if the minimum vCPU is set to 0, the instance will not start when there is no job...
It may be a good idea to check again for jobs, etc. as described in the following document.
https://repost.aws/knowledge-center/batch-jobs-termination

I tried disabling the job queue and the compute environment (I even actually deleted the job queue) and the instances were still restarting on their own! It's like the compute environment was thinking that it always had to have the instances running no matter what, even when no jobs were running. I eventually managed to stop it somehow because it finally let me edit the compute environment - I think it was when I also set "Job execution timeout" to 0 but I'm not sure?

Either way it doesn't seem very intuitive. I'm hesitant to try setting any minimum vCPU value now in case I just turned it off by blind luck!

When Minimum vCPUs is set greater than zero, a Batch Compute Environment (Comp Env) will try to ensure that number of vCPUs are available at all times, and will spin up however many EC2 instances it needs to do so. The number and type of EC2s it spins up depends on the instance types/families you have allowed in the Comp Env configuration.

Once you've set Minimum vCPUs greater than zero, it can be difficult to set it back to zero because Batch gives an error message like "Manually scaling down compute environment is not supported". To set Minimum vCPUs back to zero, you need to: 1) wait for any jobs running to finish (or terminate them), 2) "disconnect" any job-queues that are configured to use the Comp Env by changing the job queue configuration, 3) disable the Comp Env, 4) terminate any instances still running in the Comp Env, and 5) change Minimum vCPUs to zero.